Llama 3.2 90B Vision Instruct
90BMeta Llama 3.2
Flagship multimodal Llama. Requires dual 4090 or A100; vision adds ~3GB overhead.
131K
Max Context
2
Quant Variants
GGUF Q4_K_M
Best Quality
97.2%
Accuracy Retained
Quantization Variants
Per-quant VRAM, quality loss, and inference speed on RTX 4090
Measured = run on this site · Estimated = not run here: calculated, or a figure this site has not verified · Community = public reports
Similar models
Compare with Llama 3.2Llama 3.2 11B Vision Instruct
Meta Llama 3.2
Multimodal Llama with image understanding. Vision encoder adds ~2GB VRAM overhead.
Llama 3.2 3B Instruct
Meta Llama 3.2
Tiny but capable. Runs on 4GB VRAM or 8GB RAM, even on phones via llama.cpp.
Llama 3.2 1B Instruct
Meta Llama 3.2
Ultra-light Llama for mobile and embedded. Sub-2GB VRAM with Q4.
Llama 4 Scout 17B (16E)
Meta Llama 4
Meta Llama 4 Scout MoE (17B active / 109B total). Multimodal; needs ~68GB VRAM at Q4_K_M.
Running Llama 3.2 90B Vision Instruct locally
At Q4_K_M and 4K of context, Llama 3.2 90B Vision Instruct needs about 57.5 GB — 51.0 GB of weights, 1.25 GB of KV cache and a 5.2 GB activation buffer. The smallest cards in this index that clear that comfortably are 80 GB ones — 2 of them, from the A100 80G up — and 11 of the 78 cards here do. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs.
What longer context costs
Going from 4K to 32K adds about 8.75 GB, taking the total to 67.1 GB. Weights do not move with context — only the KV cache does, and it grows linearly, so this is the number to watch when planning for long documents. The model's native window is 128K; holding all of it at Q4_K_M would need about 100 GB.
Which build to download
This index only tracks Llama 3.2 90B Vision Instruct in GGUF, across 2 levels (Q4_K_M, Q3_K_M). Q4_K_M carries the lowest published perplexity loss at 2.8%. No throughput has been measured for this model on this site's hardware. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 — CUDA only, and now archived, so local use rather than serving.
Common questions
- How much VRAM does Llama 3.2 90B Vision Instruct need?
- About 57.5 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 67.1 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.
- Will Llama 3.2 90B Vision Instruct run on a 80GB GPU?
- Yes — at Q4_K_M and 4K context it needs about 57.5 GB, which leaves 22.5 GB spare on a A100 80G, or on any of the other 80 GB cards here (2 in all). That is the smallest size in this index that clears it comfortably; 11 of 78 do.
- Which quantization of Llama 3.2 90B Vision Instruct should I use?
- Q4_K_M has the lowest published quality loss (2.8%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, Q3_K_M.
Where this model fits
Sized at Q4_K_M with a 4K context window, smallest card first. Comfortable means the estimate uses at most 88% of the memory.
- Runs comfortably on
- A100 80G80GB · +22.5H100 80G80GB · +22.5Mac M2 Max 96G96GB · +14.5DGX Spark 128G128GB · +70.5Mac M3 Max 128G128GB · +38.5
- Tight but possible
- DGX Spark 64G64GB
- Just misses
- A40 48G+9.5GB overL40S 48G+9.5GB over
- Ships in
- GGUF
- Try it yourself
- Size it in the calculatorGenerate the run command
How to actually run this
Deployment guides for this model and this class of hardware.
This page's figures change when the model or the runtime does.Last updated 2026-10-20 RSS → /feed.xml